Inflation target-final 10/10/00 12:10 PM Page 80

10 Mexico’s Monetary Policy

Framework Under a
Floating Exchange
Rate Regime
Agustín G. Carstens and Alejandro M.
Werner1

The recent currency and financial crises in emerging markets have reignited
the debate on viable exchange rate regimes for small, open economies. One
common element in all of these crises was the adherence to a predetermined
exchange rate. Thus, several analysts have concluded that only under very spe-
cific and demanding conditions might there be a comfortable middle ground
between a floating exchange rate and the adoption of a common currency. In
Latin America this polarization in the choice of exchange rate regimes is
clearly represented by the different paths taken by Argentina and Mexico. After
more than four years of experience with a floating exchange rate regime, Mex-
ico provides an interesting case study for other emerging economies consider-
ing moving toward a more flexible exchange rate regime.

A Forced Transition to Floating: 1994–95

The fragilities that accumulated in the Mexican economy during the early
1990s and the negative shocks the country faced during 1994 culminated in
the balance of payments and financial crisis of December 1994. The crisis had
three different dimensions. The first was a current account deficit of signifi-
cant proportions, generated by domestic overspending and financed by short-
run capital inflows. The second was the equivalent of a run on Mexican exter-
nal liabilities, both government and private. The third was the banking crisis
that then began to unfold, which required immediate attention both to avoid
a domestic run on the banks and to pursue consistency in the macroeconomic

1The views expressed in this paper are the sole responsibility of the authors and do not neces-
sarily represent those of the institutions with which they are affiliated.

80
Inflation target-final 10/10/00 12:10 PM Page 81

Mexico’s Monetary Policy Framework Under a Floating Exchange Rate Regime 81

framework. To confront these three challenges, several measures were imple-

mented during 1995:
• Monetary and fiscal policies were tightened.
• The government negotiated a $52 billion international support package.
• The authorities implemented a series of programs to prevent a systemic
run on the banking system, combat moral hazard, minimize distortions,
and strengthen financial sector regulation and supervision.
The key to Mexico’s success in stabilizing relatively quickly was the adop-
tion of a consistent macroeconomic program immediately after the crisis. In
particular, the fiscal authority, by recognizing the costs of restructuring the
banking sector and by showing its commitment to deal with this problem with
fiscal resources, liberated monetary policy to pursue its primary goal of price
stability.

Monetary Policy Since 1995

The devaluation of the peso and the return to high inflation in 1995 seriously
damaged the credibility of the Bank of Mexico. Yet having abandoned using
the exchange rate as the economy’s nominal anchor, monetary policy had no
choice but to fill the vacuum.
First, to keep inflation in line, the Bank of Mexico established as an inter-
mediate target a ceiling on the growth of the monetary base for the year. Un-
fortunately, it soon became obvious that this simple monetary program was
not enough to stabilize inflationary expectations, the exchange rate, or infla-
tion itself. In early 1995 the rule-based monetary policy failed to perform as
expected, for several reasons:
• In a crisis scenario, the velocity of money becomes very unstable.
• The rule on growth of the monetary base was unable to keep the sudden
depreciation of the peso from affecting inflationary expectations and the
price level.
• The central bank had hardly any control over the monetary base in the
short run.
This experience led the Mexican authorities to modify their monetary pol-
icy, from one based solely on quantitative targets on monetary aggregates to
one using a combination of rules (on the behavior of the monetary base) and
discretion (by influencing the level of interest rates).
Inflation target-final 10/10/00 12:10 PM Page 82

82 Agustín G. Carstens and Alejandro M. Werner

To be able to implement the discretionary policy measures transparently,

the central bank decided to adopt a zero average reserve requirement. Under
this scheme, the Bank of Mexico established accounting periods of 28 days,
during which banks would be given incentives to post a zero average daily bal-
ance in the current accounts they hold at the central bank. Should a bank’s av-
erage daily balance become negative, the bank would have to pay a penalty in-
terest rate equivalent to twice the prevailing 28-day rate on CETES (bills issued
by the Mexican treasury) on the balance. On the other hand, a positive daily
average balance would mean that the bank was forgoing the returns it could
have obtained had it invested those funds in the market.
To meet the demand for bills and coins, the Bank of Mexico offers credit to
banks in daily auctions, so as to offset maturing credits previously granted to the
banks, movements in the account of the national treasury, and the monetary im-
pact of the central bank’s transactions in foreign currency. The central bank de-
termines the sum of credit to be auctioned each day, so that the overall net av-
erage daily balance of all current accounts held by banks at the central
bank—accumulated during the specific 28-day accounting period—will close
the day at a predetermined amount. If that amount is negative, the central bank
puts the banking system in a “short” position; if positive, the system is put in a
“long” position. It follows that if the central bank puts the system in the “short”
position, at least one credit institution will have to pay the penalty interest rate.
It should be stressed that the Bank of Mexico always supplies the credit nec-
essary to completely satisfy demand for bills and coins, even when the bank-
ing system is “short.” But in such cases a portion of the credit is supplied at a
higher interest rate, which is applied to the overdrafts in the current accounts
of one or several banks. When the system is “short,” the central bank exerts up-
ward pressure on interest rates, which can be quite significant but is primarily
the result of the signal given by the Bank of Mexico.
Since 1996, all the monetary programs that the Bank of Mexico has imple-
mented have included three main elements:
• A yearly annual inflation objective;
• A rules-defined base, together with quantitative commitments on the ac-
cumulation of net international reserves and the variation of net domes-
tic credit; and
• The possibility for the central bank to adjust its stance on monetary pol-
icy, should unexpected circumstances make it advisable (this element rep-
resents the use of discretion in monetary policy management).
The central bank tends to use the “short” mechanism, thereby adopting a
more restrictive stance, when it detects future inflationary pressures that are in-
Inflation target-final 10/10/00 12:10 PM Page 83

Mexico’s Monetary Policy Framework Under a Floating Exchange Rate Regime 83

consistent with the attainment of the inflationary target, when it is deemed nec-
essary to restore order in foreign exchange and money markets, or when infla-
tionary expectations are deemed out of line with respect to the original target.
In order to design a suitable monetary policy framework, it is essential to
identify the main determinants of inflation and understand how monetary
policy interacts with them to affect the rate of increase of the consumer price
index (CPI). There are two alternative explanations of the source of inflation-
ary pressures. In the first, the traditional monetary explanation, exogenous
shocks to the supply of money cause inflation. In the second, based on mod-
els that assume price rigidities, shocks to key prices in the economy (wages, the
exchange rate, or prices of public sector output) affect inflation directly, and
monetary policy partially accommodates these shocks. In this scenario, the de-
gree of policy accommodation is instrumental in determining the long-run
inflationary impact of the shock.
A close look at trends in the growth rate of the monetary base and inflation
during 1986–98 indicates that changes in inflation have preceded changes in
the growth rate of base money. This suggests that, during this period, ex-
ogenous movements in the money stock were not the fundamental cause of
inflation. Results of Granger causality tests show causality running both ways
between these two variables. However, these results do not quantify the influ-
ence that movements in either of these variables had on the other. To analyze
this issue in more detail, we estimated a vector error correction model that in-
corporates, as its endogenous variables, the CPI, base money, the exchange
rate, wages, and public sector prices.
The results of the variance decomposition exercises confirm that exogenous
movements in the monetary base have not been a cause of inflationary pres-
sures, but rather that the money base has accommodated inflationary shocks
coming from the exchange rate, wages, and public sector prices. These results
imply that discretionary policy measures (or reactions) to combat shocks
should be the main component of Mexico’s monetary program. So it is safe to
assume that inflationary pressures in Mexico have their origin in nonmone-
tary factors (validated ex post by the monetary authority). These include ex-
ternal shocks, which may generate a sharp depreciation of the currency;
changes in public sector prices; and wage revisions that are inconsistent with
the inflation target.
Confronted with these exogenous inflationary shocks, the central bank
must decide whether to accommodate their inflationary impact through its
monetary policy actions, and if so, whether totally or partially. To understand
the problems that the central bank faces, it is convenient to analyze first the
case where some accommodation is made to these shocks.
Inflation target-final 10/10/00 12:10 PM Page 84

84 Agustín G. Carstens and Alejandro M. Werner

Consider the case where an exogenous shock causes a sharp nominal de-
preciation of the currency. If this depreciation is perceived as permanent, it
will soon translate into increases in tradable goods prices, generating a higher
CPI. This, in turn, increases nominal demand for money. If the central bank
passively matches the demand and the supply of base money, this expansion
will be validated, and the central bank will have accommodated the increase in
money demand. An economics textbook would describe this as a once-and-
for-all adjustment in the price level, which should be accommodated.
However, in a country with a history of high inflation, the dynamics trig-
gered by such a depreciation are more complicated. The public might be led to
revise their expectations of future inflation upward, and this would lead to
rises in wages and nontradable goods prices, and thus to subsequent rounds of
exchange rate and wage adjustments.
The central bank should be able to offset at least part of the inflationary im-
pact of such an exogenous shock to the exchange rate (or to public sector
prices or contractual wages). It can do so, for example, by satisfying the daily
monetary base demand, but at an interest rate above that prevailing in the
market. This is, in fact, what happens when the Bank of Mexico puts banks in
“short” or increases the short position. The resulting higher interest rates may,
for example, partly reverse the impact of an exogenous exchange rate shock,
limiting the depreciation and moderating the adjustment of inflationary ex-
pectations. In fact, this was the Bank of Mexico’s attitude during 1998 and
1999, when it increased the short position on several occasions as additional
inflationary shocks became apparent.

Exchange Rate Policy Since 1995

The exchange rate of the Mexican peso has been floating freely since late 1994.
From time to time since then, the need has arisen for various reasons for the
authorities to intervene in the foreign exchange market. In all such cases, the
authorities have followed the rule that their intervention must be completely
transparent, and without defending a particular level of the exchange rate.
As already mentioned, one of the most important elements of the 1994–95
crisis was the run on the country’s external liabilities. On the public sector
side, the run concentrated on the now-famous tesobonos, which are short-term
dollar-denominated government securities. In the private sector, commercial
banks faced difficulties in rolling over their external liabilities. Even after the
international assistance package was approved, the resources were made avail-
able, and the adjustment program was in place, relatively large amortizations
of tesobonos and bank credits continued. It soon became obvious that if the ex-
Inflation target-final 10/10/00 12:10 PM Page 85

Mexico’s Monetary Policy Framework Under a Floating Exchange Rate Regime 85

cess demand for foreign exchange that generated this abrupt stock adjustment
in the holdings of Mexican liabilities was not satisfied by official intervention,
a sharp depreciation could ensue, even raising the risk of hyperinflation.
In addition, in the wake of the crisis, Mexico faced the need to rebuild its
international reserves. This had to be done, moreover, without affecting the
floating exchange rate and without sending any signals to the market that
might be interpreted as reflecting a desired level for the exchange rate. To ac-
complish this task, the Bank of Mexico implemented auctions of dollar put
options.
On several occasions under the floating exchange rate regime when the
peso has depreciated sharply, liquidity in the foreign exchange markets has al-
most dried up. Under such circumstances, small changes in the demand for
foreign currency have led to disproportionate depreciations of the peso. These
conditions might lead to a devaluatory spiral, which could seriously affect in-
flation and interest rates. To moderate these extreme situations, a contingent
dollar sales scheme was introduced in February 1997.
Given the macroeconomic framework that has been maintained during the
period of floating exchange rates, the exchange rate regime has not been an im-
pediment to achieving a rapid disinflation: inflation fell from 51.7 percent in
1995 to 18.6 percent in 1998. The volatility experienced by the peso during its
float, once the macroeconomic and financial crises were contained, has been
similar to that experienced by other currencies with floating exchange rates.
As a simple test of the effects of the different exchange rate regimes on in-
terest rate levels and volatilities, we can compare the behavior of interest rates
in the 1996–99 period with that observed during 1989–94, when the perfor-
mance of inflation was similar. In the current stabilization effort, under a
floating exchange rate, interest rates have usually been lower and less volatile
than during the years of the Economic Solidarity Pact (1987–94), when a pre-
determined exchange rate was in place.
In Mexico’s experience, the adoption of a floating exchange rate regime has
contributed substantially to reducing speculative pressures in financial mar-
kets. A very important feature of this regime is that it discourages short-term
capital flows, because of the risk to investors of large losses from exchange rate
fluctuations in the short run.
The flexible exchange rate also facilitates the adjustment of the real ex-
change rate toward equilibrium whenever an external shock warrants a new
equilibrium real exchange rate, without seriously harming the credibility of
the monetary authority. These movements in the real exchange rate are useful
to minimize the effects of external shocks on the economy. However, given
Mexico’s history of high inflation, inflation expectations react immediately
Inflation target-final 10/10/00 12:10 PM Page 86

86 Agustín G. Carstens and Alejandro M. Werner

when the peso depreciates. Therefore, the inflation cost of achieving the nec-
essary correction in the real exchange rate has been significantly higher than
in other countries.

Conclusion
The recent crises in financial markets underscore the importance of maintain-
ing a consistent macroeconomic framework, whatever the choice of exchange
rate regime, so as to avoid financial and balance of payments crises and achieve
long-lasting stability. It is clear that assigning multiple objectives to a single
monetary policy instrument has led to the collapse of several regimes based on
fixed or at least predetermined exchange rates. Therefore one of the most im-
portant steps that Mexico undertook after the collapse of the peso was to spell
out clearly that monetary policy was going to be focused exclusively on attain-
ing its medium-run goal of price stability. Also important was communicating
the determination that the banking sector problem was going to be addressed
by specific programs, whose cost the fiscal authority would assume.
The floating exchange rate regime has not proved to be an obstacle to Mex-
ico’s efforts to disinflate. On the contrary, it has contributed significantly to the
adjustment of the economy to external shocks and to discouraging short-term
capital inflows. Thus the floating exchange rate has become a very important
element of Mexico’s current macroeconomic policy framework.